CN105263024B - A kind of registration of HEVC video flowing zero watermarkings of anti-quantization transcoding and detection method - Google Patents

A kind of registration of HEVC video flowing zero watermarkings of anti-quantization transcoding and detection method Download PDF

Info

Publication number
CN105263024B
CN105263024B CN201510664347.0A CN201510664347A CN105263024B CN 105263024 B CN105263024 B CN 105263024B CN 201510664347 A CN201510664347 A CN 201510664347A CN 105263024 B CN105263024 B CN 105263024B
Authority
CN
China
Prior art keywords
value
org
denoted
watermark
depth
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201510664347.0A
Other languages
Chinese (zh)
Other versions
CN105263024A (en
Inventor
蒋刚毅
王静
郁梅
陈芬
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Ningbo University
Original Assignee
Ningbo University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Ningbo University filed Critical Ningbo University
Priority to CN201510664347.0A priority Critical patent/CN105263024B/en
Publication of CN105263024A publication Critical patent/CN105263024A/en
Priority to US15/045,256 priority patent/US9947065B2/en
Application granted granted Critical
Publication of CN105263024B publication Critical patent/CN105263024B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T1/00—General purpose image data processing
    • G06T1/0021—Image watermarking
    • G06T1/0028—Adaptive watermarking, e.g. Human Visual System [HVS]-based watermarking
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T1/00—General purpose image data processing
    • G06T1/0021—Image watermarking
    • G06T1/005—Robust watermarking, e.g. average attack or collusion attack resistant
    • G06T1/0057—Compression invariant watermarking
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/40—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using video transcoding, i.e. partial or full decoding of a coded input stream followed by re-encoding of the decoded output stream
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/46—Embedding additional information in the video signal during the compression process
    • H04N19/467—Embedding additional information in the video signal during the compression process characterised by the embedded information being invisible, e.g. watermarking
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2201/00—General purpose image data processing
    • G06T2201/005—Image watermarking
    • G06T2201/0053—Embedding of the watermark in the coding stream, possibly without decoding; Embedding of the watermark in the compressed domain

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Editing Of Facsimile Originals (AREA)
  • Compression Or Coding Systems Of Tv Signals (AREA)

Abstract

Registration and detection method the invention discloses a kind of HEVC video flowing zero watermarkings of anti-quantization transcoding, the register method is during using depth characteristic construction registration watermark information, depth binary conversion treatment first is carried out to the corresponding depth value of all luminance blocks in target video stream, it can be good at the robustness of reflection registration watermark by depth binary conversion treatment, so as to improve the anti-attack ability of registration watermark；Using a random matrix to being encrypted by the first watermark information matrix that the partial value after depth binary conversion treatment forms, the safety of registration watermark can be improved；The register method directly utilizes the depth characteristic of video flowing generation zero watermarking, modification initial carrier information is not needed to, the watermark transparency is not influenced, meanwhile, there is very strong robustness for attacks such as re-quantization transcoding of the quantization parameter in certain variation range and common signal processings.

Description

A kind of registration of HEVC video flowing zero watermarkings of anti-quantization transcoding and detection method
Technical field
The present invention relates to a kind of video zero watermarking technology, more particularly, to a kind of zero water of HEVC video flowings of anti-quantization transcoding The registration of print and detection method.
Background technology
Demand of the people to video gradually develops to high definition, high quality direction.To adapt to this development trend, ITU-T's The Video coding joint group of the Motion Picture Experts Group of Video Coding Experts Group and ISO/EC composition has developed high efficiency and regards jointly Frequency coding (High Efficiency Video Coding, HEVC) standard, and the formal publication in 2013, extensively should For in major video operator.However, with the maturation of multimedia and network, the acquisition and processing of video also become more Add easily, pirate phenomenon getting worse.Effectively video can be authenticated for digital watermark technology and copyright protection.Therefore, it is right The research of HEVC video flowing watermarking algorithms seems significant.
HD video data volume is huge, it is compressed for the saving costs such as bandwidth and memory space unavoidable.Based on original The water mark method of beginning video is not directly applicable compression video, therefore the watermarking algorithm of compression domain is a work for being worth research Make.At present, the watermark achievement in research in relation to HEVC compression domains is also seldom, and existing method is difficult to resist the attacks such as re-quantization transcoding And the mainly scheme of watermark embedder, it needs to be embedded in watermark by changing carrier content.However, compared to H.264/AVC, HEVC can be under similar perceived quality, and code check reduces 50% or so, and high compression rate is to above-mentioned watermark embedder algorithm in robust Property and the transparency balance bring challenge.Zero watermarking technology only need to be using the important feature of initial carrier come formation official seal Breath without changing initial carrier, can solve the HD video degradation problems because of caused by the data of change HEVC video flowings.Zero Watermark is just used in from after proposing in the content safety protection of the various information carriers such as image, video, audio.But it is directed to and regards The research of the zero watermarking algorithm of frequency compression domain is also rarely reported.
Invention content
The technical problems to be solved by the invention are to provide a kind of registration of the HEVC video flowing zero watermarkings of anti-quantization transcoding And detection method, it, can be to quantization parameter in certain variation range on the basis of modification initial carrier data are not needed to Re-quantization transcoding and the attacks such as common signal processing there is very strong robustness.
Technical solution is used by the present invention solves above-mentioned technical problem：A kind of HEVC video flowings zero of anti-quantization transcoding The register method of watermark, it is characterised in that include the following steps：
1. end _ 1, is registered in watermark, by the HEVC videos V that resolution ratio to be protected is M × NorgIt is encoded by HEVC standard Platform carries out compressed encoding, obtains target video stream, is denoted as str.binorg；And a set D is set, the initial value for enabling D is It is empty；
1. str.bin _ 2, is parsed as unit of frameorg, by str.binorgIn current f frame video frame definitions to be resolved For present frame, wherein, f is integer, and the initial value of f is 1, and 1≤f≤F, F represent str.binorgIn total frame of video frame for including Number；
1. present frame _ 3, is parsed as unit of coding tree unit, b-th of code tree to be resolved current in present frame is single Member is defined as current resolution block, wherein, b is integer, and the initial value of b is 1, and 1≤b≤B, B represent the coding included in present frame Set the total number of unit；
1. being _ 4,4 × 4 luminance block for unit using size, sequentially parse and currently parse each brightness in the block The corresponding depth value of block will currently parse the corresponding depth value of k-th of luminance block in the block and be denoted asWherein, k is integer, k Initial value for 1,1≤k≤256,
1. _ 5, enabling b=b+1, then using coding tree unit to be resolved next in present frame as current resolution block, connect Return to step 1. _ 4 to continue to execute, until all coding tree units parsing in present frame finishes, then perform step 1. _ 6, In, "=" in b=b+1 is assignment；
1. f=f+1 _ 6, is enabled, then by str.binorgMiddle next frame video frame to be resolved is then returned as present frame It returns step 1. _ 3 to continue to execute, until str.binorgIn all video frame parsing finish, then perform step 1. _ 7, wherein, f "=" in=f+1 is assignment；
1. _ 7, by str.binorgIn the corresponding depth value of all luminance blocks be sequentially stored in D, then in D The values of all elements carry out depth binary conversion treatment, obtained set is denoted as D', wherein, the element included in D it is total It counts as S, S=F × B × 256；
1. _ 8, by copyright trademark image logo of the resolution ratio for X × YorgBianry image is converted to, is denoted as logoorg'；So Afterwards by logoorg' the image of Chaotic Scrambling Cheng Xin is carried out, it is denoted as logoorg", wherein, 1≤X × Y≤F × B × 256；
1. X × Y element _ 9, is extracted from D' according to the rule of setting；Then the X that will be extracted by Z-shaped scanning form The first watermark information matrix that × Y element arranged in sequence is X × Y into a size；Simultaneously random one size of generation for X × Y and only by being worth the random matrix formed for 0 and 1 element, is denoted as Rorg；Again by each element in the first watermark information matrix Value and RorgThe value of middle corresponding element carries out exclusive or, using obtained new matrix as the first encrypted watermark information matrix, It is denoted as D*；
1. _ 10, by logoorg" in each element value and D*The value of middle corresponding element carries out exclusive or, new by what is obtained Matrix as registration watermark information, be denoted as Worg；Then request time certification authority is to WorgTimestamp is capped, when obtaining capping Between stab after registration watermark information, be denoted as Worg'；Again by Worg' be registered in the IPR databases of trusted third party's foundation, simultaneously By logoorg'、RorgWatermark detection end is sent to the rule of setting.
The step 1. _ 7 in the detailed process of depth binary conversion treatment carried out to the values of all elements in D be：
1. D _ 7a, to be decomposed into four subsets, it is denoted as G respectively0、G1、G2And G3, G0For D intermediate values be 0 all elements sequentially Extract the new set of composition, G1The all elements for being 1 for D intermediate values sequentially extract the new set of composition, G2For D Intermediate value is that 2 all elements sequentially extract the new set of composition, G3The all elements for being 3 for D intermediate values sequentially extract Come the new set formed；
1. four subsets _ 7b, are divided into two depth value sets respectively by three kinds of different packet modes, by the first point Group is denoted as C1,Wherein,Represent first depth under the first packet mode Angle value set,Represent second depth value set under the first packet mode, symbol " ∪ " is union operation symbol；By Two kinds of groupings are denoted as C2,Wherein,Represent under second of packet mode One depth value set,Represent second depth value set under second of packet mode；The third grouping is denoted as C3,Wherein,Represent first depth value set under the third packet mode,Represent second depth value set under the third packet mode；
1. _ 7c, adaptively from C1、C2And C3In select a kind of grouping, then by first depth in selected grouping The value of all elements in value set is mapped as watermark information " 0 ", and by second depth value set in selected grouping In the values of all elements be mapped as watermark information " 1 ", so far completed the depth binaryzation to the value of all elements in D Processing.
The step 1. _ 7c in adaptively from C1、C2And C3In select a kind of detailed process of grouping be：
1_1) count G0、G1、G2And G3The total number of element included in respectively, correspondence are denoted as S0、S1、S2And S3；
1_3) enable Pdiff(C1)=| P0-(P1+P2+P3) |, enable Pdiff(C2)=| (P0+P1)-(P2+P3) |, enable Pdiff(C3)= |(P0+P1+P2)-P3|, wherein, symbol " | | " it is the symbol that takes absolute value；
1_4) enable C=C2, then judge Pdiff(C) whether it is more than Pdiff(C3), if it is, enabling C=C3, then perform step Rapid 1_5)；Otherwise, keep C constant, then perform step 1_5), wherein, C=C3In "=" be assignment；
1_5) judge Pdiff(C) whether it is more than the first decision threshold T of setting, if it is, performing step 1_6)；It is no Then, keep C constant, then using C as finally selected grouping；
1_6) judge Pdiff(C) whether it is greater than or equal to Pdiff(C1), if it is, enabling C=C1, then using C as final Selected grouping；Otherwise, keep C constant, then using C as finally selected grouping, wherein, C=C1In "=" be assignment Symbol.
The step 1_5) in the value of the first decision threshold T set as 0.4.
A kind of corresponding anti-quantization transcoding of register method with the HEVC video flowing zero watermarkings of the anti-quantization transcoding HEVC video flowing zero watermarkings detection method, it is characterised in that include the following steps：
2. _ 1, at watermark detection end, by the HEVC videos V that resolution ratio to be detected is M × NdecIt is encoded by HEVC standard Platform carries out compressed encoding, obtains target video stream, is denoted as str.bindec；And a set is setIt enablesInitial value be It is empty；
2. str.bin _ 2, is parsed as unit of framedec, by str.bindecIn current f' frame video frame to be resolved determine Justice is present frame, wherein, f' is integer, and the initial value of f' is 1, and 1≤f'≤F', F' represent str.bindecIn the video that includes The totalframes of frame, F' are equal with F；
2. present frame _ 3, is parsed as unit of coding tree unit, by the b' code tree to be resolved current in present frame Unit is defined as current resolution block, wherein, b' is integer, and the initial value of b' is 1, and 1≤b'≤B', B' represent to include in present frame Coding tree unit total number；
2. being _ 4,4 × 4 luminance block for unit using size, sequentially parse and currently parse each brightness in the block The corresponding depth value of block will currently parse the corresponding depth value of k-th of luminance block in the block and be denoted asWherein, k is integer, k Initial value for 1,1≤k≤256,
2. b'=b'+1 _ 5, is enabled, then using coding tree unit to be resolved next in present frame as current resolution block, Step is then returned to 2. _ 4 to continue to execute, until all coding tree units parsing in present frame finishes, then perform step 2. _ 6, Wherein, "=" in b'=b'+1 is assignment；
2. f'=f'+1 _ 6, is enabled, then by str.bindecMiddle next frame video frame to be resolved is as present frame, then 2. _ 3 return to step continues to execute, until str.bindecIn all video frame parsing finish, then perform step 2. _ 7, wherein, "=" in f'=f'+1 is assignment；
2. _ 7, by str.bindecIn the corresponding depth value of all luminance blocks be sequentially stored inIn, it is then right In all elements value carry out depth binary conversion treatment, obtained set is denoted asWherein,In the element that includes Total number is S', S'=F' × B' × 256；
2. the rule of the setting at watermark detection end _ 8, is sent to according to watermark registration end, fromX × Y element of middle extraction； Then the second watermark information square for being X × Y into a size by the X × Y element arranged in sequence extracted by Z-shaped scanning form Battle array；The value of each element in second watermark information matrix and watermark registration end are sent to the R at watermark detection end againorgIn it is right The value of element is answered to carry out exclusive or, using obtained new matrix as the second encrypted watermark information matrix, is denoted asWherein, 1≤X×Y≤F'×B'×256；
2. _ 9, willIn each element the IPR databases established of value and trusted third party in the W that registersorg' in it is right The value of element is answered to carry out exclusive or, using obtained new matrix as copyright trademark information, is denoted as logodec”；Then to logodec” The image that chaos inverts random Cheng Xin is carried out, is denoted as logodec'；
2. _ 10, by logodec' with watermark registration end it is sent to the logo at watermark detection endorg' be compared, if the two Unanimously, then copyright authentication is completed.
The step 2. _ 7 in it is rightIn all elements value carry out depth binary conversion treatment detailed process be：
2. _ 7a, generalFour subsets are decomposed into, are denoted as respectivelyWithForIntermediate value for 0 institute The new set formed is sequentially extracted by element,ForThe all elements that intermediate value is 1 sequentially extract the new of composition Set,ForIntermediate value is that 2 all elements sequentially extract the new set of composition,ForIntermediate value be 3 it is all Element sequentially extracts the new set of composition；
2. _ 7c, adaptively fromWithIn select a kind of grouping, then by first in selected grouping The value of all elements in depth value set is mapped as watermark information " 0 ", and by second depth value in selected grouping The value of all elements in set is mapped as watermark information " 1 ", has so far completed pairIn all elements value depth Binary conversion treatment.
The step 2. _ 7c in adaptively fromWithIn select a kind of detailed process of grouping be：
2_1) countWithThe total number of element included in respectively, correspondence are denoted asWith
2_4) enableThen judgeWhether it is more thanIf it is, it enablesThen step is performed Rapid 2_5)；Otherwise, it keepsIt is constant, then perform step 2_5), wherein,In "=" be assignment；
2_5) judgeWhether second decision threshold T' of setting is more than, if it is, performing step 2_6)；Otherwise, It keepsIt is constant, then willAs finally selected grouping；
2_6) judgeWhether it is greater than or equal toIf it is, it enablesThen willAs final choosing Fixed grouping；Otherwise, it keepsIt is constant, then willAs finally selected grouping, wherein,In "=" for assignment accord with Number.
The step 2_5) in the value of the second decision threshold T' set as 0.4.
Compared with prior art, the advantage of the invention is that：
1) by counting the change in depth of coding unit before and after the re-quantization transcoding of video sequence, it is found that depth characteristic has very Good robustness, therefore the register method of the HEVC video flowing zero watermarkings of the present invention is believed using depth characteristic construction registration watermark During breath, depth binary conversion treatment first is carried out to the corresponding depth value of all luminance blocks in target video stream, is led to The robustness that depth binary conversion treatment can be good at reflection registration watermark is crossed, so as to improve the attack resistance energy of registration watermark Power.
2) register method of HEVC video flowing zero watermarkings of the invention utilizes a random matrix to by depth binaryzation First watermark information matrix of the partial value composition after reason is encrypted, and can improve the safety of registration watermark.
3) register method of HEVC video flowing zero watermarkings of the invention directly generates zero water using the depth characteristic of video flowing Print does not need to modification initial carrier information, does not influence the watermark transparency.
4) watermark can Blind extracting, the participation of original video is not needed to, and only depth part need to be decoded, without solving completely Code, computation complexity are low.
Description of the drawings
Fig. 1 a are that the overall of the register method of the present invention realizes block diagram；
Fig. 1 b are that the overall of the detection method of the present invention realizes block diagram；
Fig. 2 a are the copyright trademark image selected (resolution sizes are 205 × 650 " ningda " image)；
Fig. 2 b are the bianry image of Fig. 2 a；
Fig. 3 a are, as HEVC videos to be detected, to be obtained most not by the PartyScene cycle tests of any attack Copyright trademark image afterwards；
Fig. 3 b are so that not by the BQMall cycle tests of any attack, as HEVC videos to be detected, what is obtained is last Copyright trademark image；
Fig. 4 a are BasketballPass, ParkScene, PartyScene, Vidyo1 cycle tests in original coding amount The accuracy curve graph of watermark extracting after re-quantization attack when change parameter QP is 24；
Fig. 4 b are BasketballPass, ParkScene, PartyScene, Vidyo1 cycle tests in original coding amount The accuracy curve graph of watermark extracting after re-quantization attack when change parameter QP is 28；
Fig. 4 c are BasketballPass, ParkScene, PartyScene, Vidyo1 cycle tests in original coding amount The accuracy curve graph of watermark extracting after re-quantization attack when change parameter QP is 32；
Fig. 5 a pass through Gaussian Blur for BasketballPass, ParkScene, PartyScene, Vidyo1 cycle tests The accuracy curve graph of watermark extracting after processing and re-quantization attack；
Fig. 5 b pass through brightness deterioration for BasketballPass, ParkScene, PartyScene, Vidyo1 cycle tests The accuracy curve graph of watermark extracting after processing and re-quantization attack；
Fig. 5 c enhance for BasketballPass, ParkScene, PartyScene, Vidyo1 cycle tests by brightness The accuracy curve graph of watermark extracting after processing and re-quantization attack；
Fig. 5 d pass through sharpening operation for BasketballPass, ParkScene, PartyScene, Vidyo1 cycle tests And after re-quantization attack watermark extracting accuracy curve graph；
Fig. 5 e pass through Gaussian noise for BasketballPass, ParkScene, PartyScene, Vidyo1 cycle tests The accuracy curve graph of watermark extracting after processing and re-quantization attack.
Specific embodiment
The present invention is described in further detail below in conjunction with attached drawing embodiment.
Embodiment one：
The register method of the HEVC video flowing zero watermarkings of a kind of anti-quantization transcoding that the present embodiment proposes, is registered in watermark End, the depth characteristic and two-value for first extracting HEVC videos to be protected turn to watermark；Then, further to strengthen the safety registered Property and watermark visual effect, watermark is encrypted and with after Chaotic Scrambling copyright trademark information carry out exclusive or processing, knot Fruit is known as registering watermark；Finally, the uniqueness and authority for guarantee watermark asks the time certification mechanism of authority to believe registration Breath capping timestamp, and it is registered in intellectual property (the Intellectual Property of trusted third party's foundation Right, IPR) in database.
The overall of the register method of the HEVC video flowing zero watermarkings of the anti-quantization transcoding of the present embodiment realizes block diagram such as Fig. 1 a It is shown, specifically include following steps：
1. end _ 1, is registered in watermark, by the HEVC videos V that resolution ratio to be protected is M × NorgIt is encoded by HEVC standard Platform carries out compressed encoding, obtains target video stream, is denoted as str.binorg；And a set D is set, the initial value for enabling D is It is empty.
1. str.bin _ 2, is parsed as unit of frameorg, by str.binorgIn current f frame video frame definitions to be resolved For present frame, wherein, f is integer, and the initial value of f is 1, and 1≤f≤F, F represent str.binorgIn total frame of video frame for including Number.
1. present frame _ 3, is parsed for unit with coding tree unit (CTU), by b-th of volume to be resolved current in present frame Code tree unit is defined as current resolution block, wherein, b is integer, and the initial value of b is 1, and 1≤b≤B, B represent to include in present frame Coding tree unit total number.
1. being _ 4,4 × 4 luminance block for unit using size, sequentially parse and currently parse each brightness in the block The corresponding depth value of block will currently parse the corresponding depth value of k-th of luminance block in the block and be denoted asWherein, k is integer, k Initial value for 1,1≤k≤256,
1. _ 5, enabling b=b+1, then using coding tree unit to be resolved next in present frame as current resolution block, connect Return to step 1. _ 4 to continue to execute, until all coding tree units parsing in present frame finishes, then perform step 1. _ 6, In, "=" in b=b+1 is assignment.
1. f=f+1 _ 6, is enabled, then by str.binorgMiddle next frame video frame to be resolved is then returned as present frame It returns step 1. _ 3 to continue to execute, until str.binorgIn all video frame parsing finish, then perform step 1. _ 7, wherein, f "=" in=f+1 is assignment.
1. _ 7, by str.binorgIn the corresponding depth value of all luminance blocks be sequentially stored in D, then in D The values of all elements carry out depth binary conversion treatment, obtained set is denoted as D', wherein, the element included in D it is total It counts as S, S=F × B × 256.
In this particular embodiment, step 1. _ 7 in the tool of depth binary conversion treatment is carried out to the values of all elements in D Body process is：
1. D _ 7a, to be decomposed into four subsets, it is denoted as G respectively0、G1、G2And G3, G0For D intermediate values be 0 all elements sequentially Extract the new set of composition, G1The all elements for being 1 for D intermediate values sequentially extract the new set of composition, G2For D Intermediate value is that 2 all elements sequentially extract the new set of composition, G3The all elements for being 3 for D intermediate values sequentially extract Come the new set formed.
1. four subsets _ 7b, are divided into two depth value sets respectively by three kinds of different packet modes, by the first point Group is denoted as C1,Wherein,Represent first depth under the first packet mode Angle value set,Represent second depth value set under the first packet mode, symbol " ∪ " is union operation symbol；By Two kinds of groupings are denoted as C2,Wherein,Represent under second of packet mode One depth value set,Represent second depth value set under second of packet mode；The third grouping is denoted as C3,Wherein,Represent first depth value set under the third packet mode,Represent second depth value set under the third packet mode.
1. _ 7c, adaptively from C1、C2And C3In select a kind of grouping, then by first depth in selected grouping The value of all elements in value set is mapped as watermark information " 0 ", and by second depth value set in selected grouping In the values of all elements be mapped as watermark information " 1 ", so far completed the depth binaryzation to the value of all elements in D Processing.
In this particular embodiment, step 1. _ 7c in adaptively from C1、C2And C3In select a kind of grouping it is specific Process is：
1_1) count G0、G1、G2And G3The total number of element included in respectively, correspondence are denoted as S0、S1、S2And S3。
1_3) enable Pdiff(C1)=| P0-(P1+P2+P3) |, enable Pdiff(C2)=| (P0+P1)-(P2+P3) |, enable Pdiff(C3)= |(P0+P1+P2)-P3|, wherein, symbol " | | " it is the symbol that takes absolute value.
1_4) enable C=C2, then judge Pdiff(C) whether it is more than Pdiff(C3), if it is, enabling C=C3, then perform step Rapid 1_5)；Otherwise, keep C constant, then perform step 1_5), wherein, C=C3In "=" be assignment.
1_5) judge Pdiff(C) whether it is more than the first decision threshold T of setting, if it is, performing step 1_6)；It is no Then, keep C constant, then using C as finally selected grouping.
In the present embodiment, the value of the first decision threshold T set is 0.4.
1_6) judge Pdiff(C) whether it is greater than or equal to Pdiff(C1), if it is, enabling C=C1, then using C as final Selected grouping；Otherwise, keep C constant, then using C as finally selected grouping, wherein, C=C1In "=" be assignment Symbol.
1. _ 8, by copyright trademark image logo of the resolution ratio for X × YorgBianry image is converted to, is denoted as logoorg'；So Afterwards by logoorg' the image of Chaotic Scrambling Cheng Xin is carried out, it is denoted as logoorg", wherein, 1≤X × Y≤F × B × 256.
In the present embodiment, logoorg" ningda " image that resolution sizes are 205 × 650 is selected, as shown in Figure 2 a, The image is coloured image, and there are three channels, then convert the image to bianry image logoorg', as shown in Figure 2 b.
In the present embodiment, Chaotic Scrambling is by bianry image logoorg' in data into row stochastic out of order operation.
1. X × Y element _ 9, is extracted from D' according to the rule of setting；Then the X that will be extracted by Z-shaped scanning form The first watermark information matrix that × Y element arranged in sequence is X × Y into a size；Simultaneously random one size of generation for X × Y and only by being worth the random matrix formed for 0 and 1 element, is denoted as Rorg；Again by each element in the first watermark information matrix Value and RorgThe value of middle corresponding element carries out exclusive or, using obtained new matrix as the first encrypted watermark information matrix, It is denoted as D*。
In the present embodiment, the rule set such as can be the preceding X × Y element taken in D' as the rule of sets itself, Can also take the rear X × Y element in D' or intermediate X × Y element or interval in D' is taken to take D' In X × Y element, can also be other various modes.
In the present embodiment, random matrix can be used existing tool and directly generate, such as using MATLAB or c++.
1. _ 10, by logoorg" in each element value and D*The value of middle corresponding element carries out exclusive or, new by what is obtained Matrix as registration watermark information, be denoted as Worg；Then request time certification authority is to WorgTimestamp is capped, when obtaining capping Between stab after registration watermark information, be denoted as Worg'；Again by Worg' it is registered in the IPR (Intellectual that trusted third party establishes Property Right) in database, while by logoorg'、RorgWatermark detection end is sent to the rule of setting.
In the present embodiment, to registering watermark information WorgThe timestamp of capping is provided by the time certification mechanism of authority Registion time prove, can effectively confirm the ownership of works right.
Embodiment two：
The present embodiment proposes a kind of registration of the HEVC video flowing zero watermarkings of the anti-quantization transcoding proposed with embodiment one The detection method of the HEVC video flowing zero watermarkings of the corresponding anti-quantization transcoding of method, it is overall to realize block diagram as shown in Figure 1 b, Its at watermark detection end, first extract depth information in video to be detected and by binaryzation, and carried out with the key preserved Encryption；The watermark information with IPR database registrations makees xor operation again；Finally, chaos is carried out to the result after operation and inverts unrest, Adjudicate copyright ownership.
The overall of the detection method of the HEVC video flowing zero watermarkings of the anti-quantization transcoding of the present embodiment realizes block diagram such as Fig. 1 b It is shown, specifically include following steps：
2. _ 1, at watermark detection end, by the HEVC videos V that resolution ratio to be detected is M × NdecIt is encoded by HEVC standard Platform carries out compressed encoding, obtains target video stream, is denoted as str.bindec；And a set is setIt enablesInitial value be It is empty.
In the present embodiment, it is desirable that HEVC videos VdecWith HEVC videos VorgResolution ratio it is consistent.
2. str.bin _ 2, is parsed as unit of framedec, by str.bindecIn current f' frame video frame to be resolved determine Justice is present frame, wherein, f' is integer, and the initial value of f' is 1, and 1≤f'≤F', F' represent str.bindecIn the video that includes The totalframes of frame, F' are equal with F.
2. present frame _ 3, is parsed for unit with coding tree unit (CTU), by the b' volume to be resolved current in present frame Code tree unit is defined as current resolution block, wherein, b' is integer, and the initial value of b' is 1, and 1≤b'≤B', B' are represented in present frame Comprising coding tree unit total number.
2. being _ 4,4 × 4 luminance block for unit using size, sequentially parse and currently parse each brightness in the block The corresponding depth value of block will currently parse the corresponding depth value of k-th of luminance block in the block and be denoted asWherein, k is integer, k Initial value for 1,1≤k≤256,
2. b'=b'+1 _ 5, is enabled, then using coding tree unit to be resolved next in present frame as current resolution block, Step is then returned to 2. _ 4 to continue to execute, until all coding tree units parsing in present frame finishes, then perform step 2. _ 6, Wherein, "=" in b'=b'+1 is assignment.
2. f'=f'+1 _ 6, is enabled, then by str.bindecMiddle next frame video frame to be resolved is as present frame, then 2. _ 3 return to step continues to execute, until str.bindecIn all video frame parsing finish, then perform step 2. _ 7, wherein, "=" in f'=f'+1 is assignment.
2. _ 7, by str.bindecIn the corresponding depth value of all luminance blocks be sequentially stored inIn, it is then right In all elements value carry out depth binary conversion treatment, obtained set is denoted asWherein,In the element that includes Total number is S', S'=F' × B' × 256.
In this particular embodiment, step 2. _ 7 in it is rightIn all elements value carry out depth binary conversion treatment tool Body process is：
2. _ 7a, generalFour subsets are decomposed into, are denoted as respectivelyWithForIntermediate value for 0 institute The new set formed is sequentially extracted by element,ForThe all elements that intermediate value is 1 sequentially extract the new of composition Set,ForIntermediate value is that 2 all elements sequentially extract the new set of composition,ForIntermediate value be 3 it is all Element sequentially extracts the new set of composition.
2. _ 7c, adaptively fromWithIn select a kind of grouping, then by first in selected grouping The value of all elements in depth value set is mapped as watermark information " 0 ", and by second depth value in selected grouping The value of all elements in set is mapped as watermark information " 1 ", has so far completed pairIn all elements value depth Binary conversion treatment.
In this particular embodiment, step 2. _ 7c in adaptively fromWithIn select a kind of tool of grouping Body process is：
2_1) countWithThe total number of element included in respectively, correspondence are denoted asWith
2_4) enableThen judgeWhether it is more thanIf it is, it enablesThen step is performed Rapid 2_5)；Otherwise, it keepsIt is constant, then perform step 2_5), wherein,In "=" be assignment.
2_5) judgeWhether second decision threshold T' of setting is more than, if it is, performing step 2_6)；It is no Then, it keepsIt is constant, then willAs finally selected grouping.
In the present embodiment, the value of the second decision threshold T' set is 0.4.
2_6) judgeWhether it is greater than or equal toIf it is, it enablesThen willAs final choosing Fixed grouping；Otherwise, it keepsIt is constant, then willAs finally selected grouping, wherein,In "=" for assignment accord with Number.
2. the rule of the setting at watermark detection end _ 8, is sent to according to watermark registration end, fromX × Y element of middle extraction； Then the second watermark information square for being X × Y into a size by the X × Y element arranged in sequence extracted by Z-shaped scanning form Battle array；The value of each element in second watermark information matrix and watermark registration end are sent to the R at watermark detection end againorgIn it is right The value of element is answered to carry out exclusive or, using obtained new matrix as the second encrypted watermark information matrix, is denoted asWherein, 1≤X×Y≤F'×B'×256。
In the present embodiment, fromThe rule that the middle rule for extracting the second watermark information matrix is extracted with watermark registration end Unanimously, the size of X and Y is constant.
2. _ 9, willIn each element IPR (the Intellectual Property that establish of value and trusted third party Right) the W registered in databaseorg' in corresponding element value carry out exclusive or, using obtained new matrix as copyright trade mark letter Breath, is denoted as logodec”；Then to logodec" image that chaos inverts random Cheng Xin is carried out, it is denoted as logodec'。
2. _ 10, by logodec' with watermark registration end it is sent to the logo at watermark detection endorg' be compared, if the two Unanimously, then copyright authentication is completed.
The register method of the invention for verification and the validity and feasibility of corresponding detection method, to the method for the present invention It is tested.
Using 4 different resolutions and the different cycle tests BasketballPass, ParkScene of Texture complication, PartyScene, Vidyo1, resolution ratio corresponds to 416 × 240,1920 × 1080,832 × 480,1280 × 720, in this implementation Using standard testing encoding platform HM15.0 in example, using full frame interior coding, coding frame number is 100 frames.Separately below from zero water The robustness and type signal of re-quantization transcoding attack of the independence, coded quantization parameter QP of print in certain variation range Processing combines correctness PC (Percent of these three aspects of the robustness of re-quantization transcoding attack according to watermark extracting Correct) validity of the method for the present invention is evaluated, the formula of PC is：Wherein, w Watermark information after the depth binaryzation of k-th of luminance block in f frames after the original HEVC Video codings of (f, k) expression, After w ' (f, k) represents the depth binaryzation of k-th of luminance block in the f frames after the recodification of the HEVC videos after attack Watermark information, F are the totalframes for the video frame that video sequence includes, and K is the total number of luminance block that every frame video frame includes, symbol Number " ⊙ " be with or functional symbol.
1) independence of zero watermarking
Zero watermarking is the distinctive characterization of carrier, therefore the zero watermarking of different video sequences should be mutual indepedent.With The independence of zero watermarking, point of the two cycle tests are tested for PartyScene cycle tests and BQMall cycle tests Resolution is all 832 × 480.PartyScene cycle tests and BQMall cycle tests register method according to the invention are obtained Register watermark information.Fig. 3 a give using the PartyScene cycle tests not by any attack and are regarded as HEVC to be detected Frequently, the last copyright trademark image obtained；Fig. 3 b give using not by the BQMall cycle tests of any attack as to be checked The HEVC videos of survey, obtained last copyright trademark image.From the point of view of obtained last copyright trademark image, when to be detected HEVC videos as original licensed video when, can completely obtain original copyright trademark image；When to be detected When HEVC videos and different original licensed video, it is impossible to original copyright trademark image is obtained, it is different so as to illustrate The zero watermarking of video sequence should be mutual indepedent.
2) resist the robustness of re-quantization transcoding
Quantify transcoding to destroy compression domain watermark under the premise of certain video quality loss is ensured, be most common pressure One of contracting video watermarking attack method is to carry out compressed encoding again to video sequence.The original coding quantization of cycle tests Parameter QP takes 24,28,32 respectively.Fig. 4 a give BasketballPass, ParkScene, PartyScene, Vidyo1 test The accuracy curve graph of watermark extracting after re-quantization attack of the sequence when original coding quantization parameter QP is 24；Fig. 4 b give BasketballPass, ParkScene, PartyScene, Vidyo1 cycle tests are when original coding quantization parameter QP is 28 Re-quantization attack after watermark extracting accuracy curve graph；Fig. 4 c give BasketballPass, ParkScene, PartyScene, Vidyo1 cycle tests original coding quantization parameter QP be 32 when re-quantization attack after watermark extracting just True rate curve graph.The accuracy curve of watermark extracting in Fig. 4 a to Fig. 4 c illustrates the accuracy of watermark with re-quantization transcoding Different original coding quantization parameter QP situation of change, as can be seen that the watermark of the method for the present invention carries from Fig. 4 a to Fig. 4 c Accuracy PC is taken more than 85%, illustrates that the method for the present invention can resist re-quantization turn in the range of certain coded quantization parameter QP Code attack；Accuracy variation is little in the range of [QP-3, QP], substantially all more than 95%；In the range of [QP+1, QP+3] just True rate is substantially reduced, and minimum 85.34%, increase original coding quantization parameter QP is primarily due to coding unit size decision When rate distortion costs influence it is bigger.
3) antinoise signal processing combines the robustness of re-quantization transcoding
Through various signal processings with reference to the test of the robustness after the re-quantization transcoding attack of identical coded quantization parameter QP As a result as shown in Fig. 5 a to Fig. 5 e.Fig. 5 a give BasketballPass, ParkScene, PartyScene, Vidyo1 test The accuracy curve graph of sequence watermark extracting after Gaussian Blur processing and re-quantization attack；Fig. 5 b give BasketballPass, ParkScene, PartyScene, Vidyo1 cycle tests by brightness deterioration handle and it is re-quantization The accuracy curve graph of watermark extracting after attack；Fig. 5 c give BasketballPass, ParkScene, PartyScene, The accuracy curve graph of Vidyo1 cycle tests watermark extracting after brightness enhances processing and re-quantization attack；Fig. 5 d are provided BasketballPass, ParkScene, PartyScene, Vidyo1 cycle tests is by sharpening operation and re-quantization attacks Hit the accuracy curve graph of rear watermark extracting；Fig. 5 e give BasketballPass, ParkScene, PartyScene, The accuracy curve graph of Vidyo1 cycle tests watermark extracting after Gaussian noise processing and re-quantization attack.From Fig. 5 a and In Fig. 5 d as can be seen that under a degree of filtering process, accuracy can reach more than 93%, show that the method for the present invention can To resist a degree of filtering process；From Fig. 5 b and Fig. 5 c as can be seen that under a degree of brightness attack, accuracy More than 93% can be reached, show that the method for the present invention can resist a degree of brightness attack；As can be seen that water from Fig. 5 e The PC of extraction is printed more than 88%, shows that the method for the present invention can resist a degree of attacked by noise.

Claims (8)

1. a kind of register method of the HEVC video flowing zero watermarkings of anti-quantization transcoding, it is characterised in that include the following steps：
1. end _ 1, is registered in watermark, by the HEVC videos V that resolution ratio to be protected is M × NorgPass through HEVC standard encoding platform Compressed encoding is carried out, target video stream is obtained, is denoted as str.binorg；And a set D is set, the initial value for enabling D is sky；
1. str.bin _ 2, is parsed as unit of frameorg, by str.binorgIn current f frame video frame to be resolved be defined as working as Previous frame, wherein, f is integer, and the initial value of f is 1, and 1≤f≤F, F represent str.binorgIn the totalframes of video frame that includes；
1. _ 3, parsing present frame as unit of coding tree unit, b-th of coding tree unit to be resolved current in present frame is determined Justice is current resolution block, wherein, b is integer, and the initial value of b is 1, and 1≤b≤B, B represent that the code tree included in present frame is single The total number of member；
1. being _ 4,4 × 4 luminance block for unit using size, sequentially parse and currently parse each luminance block pair in the block The depth value answered will currently parse the corresponding depth value of k-th of luminance block in the block and be denoted asWherein, k is integer, and k's is first Initial value is 1,1≤k≤256,
1. _ 5, enabling b=b+1, then using coding tree unit to be resolved next in present frame as current resolution block, then return Step is returned 1. _ 4 to continue to execute, until all coding tree units parsing in present frame finishes, then perform step 1. _ 6, wherein, b "=" in=b+1 is assignment；
1. f=f+1 _ 6, is enabled, then by str.binorgMiddle next frame video frame to be resolved is then returned to step as present frame Suddenly it 1. _ 3 continues to execute, until str.binorgIn all video frame parsing finish, then perform step 1. _ 7, wherein, f=f+1 In "=" be assignment；
1. _ 7, by str.binorgIn the corresponding depth value of all luminance blocks be sequentially stored in D, then to the institute in D The value for having element carries out depth binary conversion treatment, and obtained set is denoted as D', wherein, the total number of the element included in D is S, S=F × B × 256；
1. _ 8, by copyright trademark image logo of the resolution ratio for X × YorgBianry image is converted to, is denoted as logoorg'；Then will logoorg' the image of Chaotic Scrambling Cheng Xin is carried out, it is denoted as logoorg", wherein, 1≤X × Y≤F × B × 256；
1. X × Y element _ 9, is extracted from D' according to the rule of setting；Then X × the Y that will be extracted by Z-shaped scanning form The first watermark information matrix that a element arranged in sequence is X × Y into a size；Simultaneously random one size of generation for X × Y and Only by being worth the random matrix formed for 0 and 1 element, it is denoted as Rorg；Again by the value of each element in the first watermark information matrix With RorgThe value of middle corresponding element carries out exclusive or, using obtained new matrix as the first encrypted watermark information matrix, is denoted as D*；
1. _ 10, by logoorg" in each element value and D*The value of middle corresponding element carries out exclusive or, the new matrix that will be obtained As registration watermark information, it is denoted as Worg；Then request time certification authority is to WorgTimestamp is capped, after obtaining capping timestamp Registration watermark information, be denoted as Worg'；Again by Worg' be registered in the IPR databases of trusted third party's foundation, simultaneously will logoorg'、RorgWatermark detection end is sent to the rule of setting.
2. a kind of register method of the HEVC video flowing zero watermarkings of anti-quantization transcoding according to claim 1, feature exist In the step 1. _ 7 in the detailed process of depth binary conversion treatment carried out to the values of all elements in D be：
1. D _ 7a, to be decomposed into four subsets, it is denoted as G respectively0、G1、G2And G3, G0The all elements for being 0 for D intermediate values sequentially extract The new set out formed, G1The all elements for being 1 for D intermediate values sequentially extract the new set of composition, G2For D intermediate values All elements for 2 sequentially extract the new set of composition, G3The all elements for being 3 for D intermediate values sequentially extract group Into new set；
1. four subsets _ 7b, are divided into two depth value sets respectively by three kinds of different packet modes, the first grouping is remembered For C1, C1={ F1 1=G0,F2 1=G1∪G2∪G3, wherein, F1 1Represent first depth value collection under the first packet mode It closes, F2 1Represent second depth value set under the first packet mode, symbol " ∪ " is union operation symbol；By second point Group is denoted as C2, C2={ F1 2=G0∪G1,F2 2=G2∪G3, wherein, F1 2Represent first depth value under second of packet mode Set, F2 2Represent second depth value set under second of packet mode；The third grouping is denoted as C3, C3={ F1 3=G0 ∪G1∪G2,F2 3=G3, wherein, F1 3Represent first depth value set under the third packet mode, F2 3Represent the third point Second depth value set under prescription formula；
1. _ 7c, adaptively from C1、C2And C3In select a kind of grouping, then by first depth value collection in selected grouping The value of all elements in conjunction is mapped as watermark information " 0 ", and will be in second depth value set in selected grouping The value of all elements is mapped as watermark information " 1 ", has so far completed at the depth binaryzation to the value of all elements in D Reason.
3. a kind of register method of the HEVC video flowing zero watermarkings of anti-quantization transcoding according to claim 2, feature exist In the step 1. _ 7c in adaptively from C1、C2And C3In select a kind of detailed process of grouping be：
1_1) count G0、G1、G2And G3The total number of element included in respectively, correspondence are denoted as S0、S1、S2And S3；
1_2) calculate S0、S1、S2And S3It is corresponding to be denoted as P respectively with the ratio of S0、P1、P2And P3,
1_3) enable Pdiff(C1)=| P0-(P1+P2+P3) |, enable Pdiff(C2)=| (P0+P1)-(P2+P3) |, enable Pdiff(C3)=| (P0 +P1+P2)-P3|, wherein, symbol " | | " it is the symbol that takes absolute value；
1_4) enable C=C2, then judge Pdiff(C) whether it is more than Pdiff(C3), if it is, enabling C=C3, then perform step 1_ 5)；Otherwise, keep C constant, then perform step 1_5), wherein, C=C3In "=" be assignment；
1_5) judge Pdiff(C) whether it is more than the first decision threshold T of setting, if it is, performing step 1_6)；Otherwise, it keeps C is constant, then using C as finally selected grouping；
1_6) judge Pdiff(C) whether it is greater than or equal to Pdiff(C1), if it is, enabling C=C1, then using C as final selected Grouping；Otherwise, keep C constant, then using C as finally selected grouping, wherein, C=C1In "=" be assignment.
4. a kind of register method of the HEVC video flowing zero watermarkings of anti-quantization transcoding according to claim 3, feature exist In the step 1_5) in the value of the first decision threshold T set as 0.4.
5. a kind of corresponding anti-amount of register method with the HEVC video flowing zero watermarkings of anti-quantization transcoding described in claim 1 Change the detection method of the HEVC video flowing zero watermarkings of transcoding, it is characterised in that include the following steps：
2. _ 1, at watermark detection end, by the HEVC videos V that resolution ratio to be detected is M × NdecPass through HEVC standard encoding platform Compressed encoding is carried out, target video stream is obtained, is denoted as str.bindec；And a set is setIt enablesInitial value for sky；
2. str.bin _ 2, is parsed as unit of framedec, by str.bindecIn current f' frame video frame to be resolved be defined as Present frame, wherein, f' is integer, and the initial value of f' is 1, and 1≤f'≤F', F' represent str.bindecIn the video frame that includes Totalframes, F' are equal with F；
2. present frame _ 3, is parsed as unit of coding tree unit, by the b' coding tree unit to be resolved current in present frame Current resolution block is defined as, wherein, b' is integer, and the initial value of b' is 1, and 1≤b'≤B', B' represent the volume included in present frame The total number of code tree unit；
2. being _ 4,4 × 4 luminance block for unit using size, sequentially parse and currently parse each luminance block pair in the block The depth value answered will currently parse the corresponding depth value of k-th of luminance block in the block and be denoted asWherein, k is integer, and k's is first Initial value is 1,1≤k≤256,
2. b'=b'+1 _ 5, is enabled, then using coding tree unit to be resolved next in present frame as current resolution block, then 2. _ 4 return to step continues to execute, until all coding tree units parsing in present frame finishes, then perform step 2. _ 6, In, "=" in b'=b'+1 is assignment；
2. f'=f'+1 _ 6, is enabled, then by str.bindecMiddle next frame video frame to be resolved is then returned to as present frame 2. _ 3 step continues to execute, until str.bindecIn all video frame parsing finish, then perform step 2. _ 7, wherein, f'= "=" in f'+1 is assignment；
2. _ 7, by str.bindecIn the corresponding depth value of all luminance blocks be sequentially stored inIn, it is then rightIn The value of all elements carries out depth binary conversion treatment, and obtained set is denoted asWherein,In include element it is total It counts as S', S'=F' × B' × 256；
2. the rule of the setting at watermark detection end _ 8, is sent to according to watermark registration end, fromX × Y element of middle extraction；Then The the second watermark information matrix for being X × Y into a size by the X × Y element arranged in sequence extracted by Z-shaped scanning form； The value of each element in second watermark information matrix and watermark registration end are sent to the R at watermark detection end againorgMiddle corresponding element The value of element carries out exclusive or, using obtained new matrix as the second encrypted watermark information matrix, is denoted asWherein, 1≤X ×Y≤F'×B'×256；
2. _ 9, willIn each element the IPR databases established of value and trusted third party in the W that registersorg' in corresponding element The value of element carries out exclusive or, using obtained new matrix as copyright trademark information, is denoted as logodec”；Then to logodec" carry out Chaos inverts the image of random Cheng Xin, is denoted as logodec'；
2. _ 10, by logodec' with watermark registration end it is sent to the logo at watermark detection endorg' be compared, if the two one It causes, then completes copyright authentication.
6. a kind of detection method of the HEVC video flowing zero watermarkings of anti-quantization transcoding according to claim 5, feature exist In the step 2. _ 7 in it is rightIn all elements value carry out depth binary conversion treatment detailed process be：
2. _ 7a, generalFour subsets are decomposed into, are denoted as respectively With ForIntermediate value is 0 all elements The new set of composition is sequentially extracted,ForIntermediate value is that 1 all elements sequentially extract the new collection of composition It closes,ForIntermediate value is that 2 all elements sequentially extract the new set of composition,ForIntermediate value is 3 all members Element sequentially extracts the new set of composition；
2. four subsets _ 7b, are divided into two depth value sets respectively by three kinds of different packet modes, the first grouping is remembered ForWherein, first depth value under the first packet mode is represented Set, represents second depth value set under the first packet mode, and symbol " ∪ " is union operation symbol；By second Kind grouping is denoted asWherein, the under second of packet mode is represented One depth value set represents second depth value set under second of packet mode；The third grouping is denoted asWherein, it represents first depth value set under the third packet mode, represents second depth value set under the third packet mode；
2. _ 7c, adaptively from WithIn select a kind of grouping, then by first depth value in selected grouping The value of all elements in set is mapped as watermark information " 0 ", and will be in second depth value set in selected grouping The values of all elements be mapped as watermark information " 1 ", so far completed pairIn all elements value depth binaryzation Processing.
7. a kind of detection method of the HEVC video flowing zero watermarkings of anti-quantization transcoding according to claim 6, feature exist In the step 2. _ 7c in adaptively from WithIn select a kind of detailed process of grouping be：
2_1) count WithThe total number of element included in respectively, correspondence are denoted as With
2_2) calculate and respectively with the ratio of S', it is corresponding be denoted as with
2_3) enableIt enablesIt enablesWherein, symbol " | | " is the symbol that takes absolute value；
2_4) enableThen judgeWhether it is more thanIf it is, it enablesThen step is performed 2_5)；Otherwise, it keepsIt is constant, then perform step 2_5), wherein,In "=" be assignment；
2_5) judgeWhether second decision threshold T' of setting is more than, if it is, performing step 2_6)；Otherwise, it protects It holdsIt is constant, then willAs finally selected grouping；
2_6) judgeWhether it is greater than or equal toIf it is, it enablesThen willAs final selected Grouping；Otherwise, it keepsIt is constant, then willAs finally selected grouping, wherein,In "=" for assignment accord with Number.
8. a kind of detection method of the HEVC video flowing zero watermarkings of anti-quantization transcoding according to claim 7, feature exist In the step 2_5) in the value of the second decision threshold T' set as 0.4.
CN201510664347.0A 2015-10-15 2015-10-15 A kind of registration of HEVC video flowing zero watermarkings of anti-quantization transcoding and detection method Active CN105263024B (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN201510664347.0A CN105263024B (en) 2015-10-15 2015-10-15 A kind of registration of HEVC video flowing zero watermarkings of anti-quantization transcoding and detection method
US15/045,256 US9947065B2 (en) 2015-10-15 2016-02-17 Zero-watermarking registration and detection method for HEVC video streaming against requantization transcoding

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201510664347.0A CN105263024B (en) 2015-10-15 2015-10-15 A kind of registration of HEVC video flowing zero watermarkings of anti-quantization transcoding and detection method

Publications (2)

Publication Number Publication Date
CN105263024A CN105263024A (en) 2016-01-20
CN105263024B true CN105263024B (en) 2018-06-29

Family

ID=55102504

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201510664347.0A Active CN105263024B (en) 2015-10-15 2015-10-15 A kind of registration of HEVC video flowing zero watermarkings of anti-quantization transcoding and detection method

Country Status (2)

Country Link
US (1) US9947065B2 (en)
CN (1) CN105263024B (en)

Families Citing this family (19)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105263024B (en) * 2015-10-15 2018-06-29 宁波大学 A kind of registration of HEVC video flowing zero watermarkings of anti-quantization transcoding and detection method
US10237608B2 (en) * 2016-09-13 2019-03-19 Facebook, Inc. Systems and methods for evaluating synchronization between content streams
CN107274331B (en) * 2017-06-13 2018-03-13 重庆第二师范学院 Robust watermarking embedding grammar and device for data flow
MX2020001303A (en) * 2017-08-10 2020-03-09 Sony Corp Transmission device, transmission method, reception device, and reception method.
CN108768606B (en) * 2018-05-10 2021-09-17 深圳市汇星数字技术有限公司 Streaming media real-time encryption method and system based on video transmission
US10937117B2 (en) 2018-07-20 2021-03-02 Comcast Cable Communications, Llc Collusion attack prevention
CN109064375B (en) * 2018-07-23 2023-05-09 北京信息科技大学 Zero watermark-based large data property identification method and system
CN111225277A (en) * 2018-11-27 2020-06-02 北京达佳互联信息技术有限公司 Transcoding method, transcoding device and computer readable storage medium
CN111310215B (en) * 2020-02-26 2020-12-11 海南大学 Multilayer digital veil design method for image content safety and privacy protection
CN111984942B (en) * 2020-07-23 2023-10-27 西安理工大学 Robust video zero watermarking method based on polar complex exponential transformation and residual neural network
CN112561771B (en) * 2020-12-08 2022-03-04 齐鲁工业大学 Robust zero watermark method based on complementary features
CN114630147B (en) * 2020-12-11 2024-03-19 上海哔哩哔哩科技有限公司 Video processing method and device
US20220191522A1 (en) * 2020-12-11 2022-06-16 Shanghai Bilibili Technology Co., Ltd. Video processing method and device
CN113012020B (en) * 2021-04-23 2023-07-21 中国科学院空间应用工程与技术中心 Image watermarking method, system and electronic equipment
CN113935346B (en) * 2021-10-12 2022-06-21 南通大学 Commodity anti-counterfeiting code generation method based on trademark picture scrambling encryption
CN114897659B (en) * 2022-05-09 2023-12-29 南京师范大学 Vector geographic data zero watermark generation method and zero watermark information detection method
CN115311119B (en) * 2022-10-09 2022-12-23 中国民航大学 Three-dimensional image zero watermark embedding and extracting method capable of resisting geometric attack
CN117241104B (en) * 2023-11-07 2024-01-16 中南大学 Zero watermark method and system for DIBR-3D video copyright protection
CN117635408A (en) * 2023-11-22 2024-03-01 南京财经大学 Copyright protection-oriented image zero-watermark method, device and medium

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH11272299A (en) * 1998-03-23 1999-10-08 Toyo Commun Equip Co Ltd Method for embedding watermark bit during voice encoding
CN101178805A (en) * 2007-12-12 2008-05-14 北京航空航天大学 Three-dimensional grid digital blind watermark method based on Octree encode
CN101908201A (en) * 2010-06-28 2010-12-08 宁波大学 Robust digital image adaptive zero-watermarking method
CN102314669A (en) * 2011-09-13 2012-01-11 海南大学 DCT (discrete cosine transform)-based anti-geometric-attack zero-digital-watermarking method for medical image
CN102682418A (en) * 2012-05-11 2012-09-19 宁波大学 Method for embedding and extracting multiple zero watermarks of digital image
CN103996165A (en) * 2014-05-30 2014-08-20 东北大学 Digital image zero watermark embedding and extracting method based on compressed sensing characteristics

Family Cites Families (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6947571B1 (en) * 1999-05-19 2005-09-20 Digimarc Corporation Cell phones with optical capabilities, and related applications
US7406214B2 (en) * 1999-05-19 2008-07-29 Digimarc Corporation Methods and devices employing optical sensors and/or steganography
JP2009508393A (en) * 2005-09-09 2009-02-26 トムソン ライセンシング Video watermark insertion
JP2009508395A (en) * 2005-09-09 2009-02-26 トムソン ライセンシング Coefficient modification for video watermark insertion
CN101681499B (en) * 2007-06-14 2013-04-24 汤姆逊许可证公司 Modifying a coded bitstream
US8401223B2 (en) * 2008-10-20 2013-03-19 Virginia Venture Industries, Llc Embedding and decoding three-dimensional watermarks into stereoscopic images
CN105263024B (en) * 2015-10-15 2018-06-29 宁波大学 A kind of registration of HEVC video flowing zero watermarkings of anti-quantization transcoding and detection method
CN106162195B (en) * 2016-07-05 2018-04-17 宁波大学 A kind of 3D HEVC deep video information concealing methods based on single depth frame mode
CN107040787B (en) * 2017-03-30 2019-08-02 宁波大学 A kind of 3D-HEVC inter-frame information hidden method of view-based access control model perception

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH11272299A (en) * 1998-03-23 1999-10-08 Toyo Commun Equip Co Ltd Method for embedding watermark bit during voice encoding
CN101178805A (en) * 2007-12-12 2008-05-14 北京航空航天大学 Three-dimensional grid digital blind watermark method based on Octree encode
CN101908201A (en) * 2010-06-28 2010-12-08 宁波大学 Robust digital image adaptive zero-watermarking method
CN102314669A (en) * 2011-09-13 2012-01-11 海南大学 DCT (discrete cosine transform)-based anti-geometric-attack zero-digital-watermarking method for medical image
CN102682418A (en) * 2012-05-11 2012-09-19 宁波大学 Method for embedding and extracting multiple zero watermarks of digital image
CN103996165A (en) * 2014-05-30 2014-08-20 东北大学 Digital image zero watermark embedding and extracting method based on compressed sensing characteristics

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
《抗重量化转码的H.264/AVC视频水印算法》;徐达文，王让定;《电子与信息学报》;20130515(第5期);全文 *
H.264 Video Standard Based Zero Watermarking Technology;Gui Feng，Kai Huang;《Anti-Counterfeiting，Security and Identification，2013 IEEE International Conference on》;20140605;全文 *
Information hiding in HEVC standard using adaptive coding block size decision;Y Tew，KS Wong;《Image Processing，2014 IEEE International Conference on》;20150129;全文 *

Also Published As

Publication number Publication date
US9947065B2 (en) 2018-04-17
CN105263024A (en) 2016-01-20
US20170109858A1 (en) 2017-04-20

Similar Documents

Publication Publication Date Title
CN105263024B (en) A kind of registration of HEVC video flowing zero watermarkings of anti-quantization transcoding and detection method
CN107040787B (en) A kind of 3D-HEVC inter-frame information hidden method of view-based access control model perception
CN101472161B (en) Method, device and system for embedding and removing watermark
Chang et al. A reversible data hiding scheme for VQ indices using locally adaptive coding
CN104581176B (en) The insertion of H.264/AVC compression domain robust video watermark and extracting method without frame in error drift
Kelash et al. Hiding data in video sequences using steganography algorithms
Liu Colour image watermarking for tamper proofing and pattern-based recovery
CN104093029B (en) A kind of Video Encryption Algorithm based on new Spatiotemporal Chaotic Systems
Mstafa et al. A new video steganography algorithm based on the multiple object tracking and Hamming codes
CN106162195A (en) A kind of 3D HEVC deep video information concealing method based on single depth frame internal schema
CN107222751B (en) 3D-HEVC deep video information concealing method based on multi-view point video feature
CN103136718A (en) Anti-geometric-transformation practical color image digital watermark technology
Chang et al. A block-based watermarking scheme for image tamper detection and self-recovery
Kurup et al. Data hiding scheme based on octagon shaped shell
CN105657431A (en) Watermarking algorithm based on DCT domain of video frame
CN106780278B (en) Self-recovery image encryption and decryption method combining zero watermark and block classification fusion
CN104902281B (en) A kind of information concealing method of the HEVC videos based on Hamming code+1
Dhaka et al. A novel algorithm for image steganography based on effective channel selection technique
Singh et al. Image security using DES and RNS with reversible watermarking
Li et al. Semi-fragile watermarking for image tamper localization and self-recovery
Fallahpour et al. Flexible image watermarking in JPEG domain
CN106851436A (en) Video point shape Fragile Watermark Algorithm in DCT domain
CN110189241B (en) Block mean value-based anti-printing noise image watermarking method
CN113938573A (en) Computational ghost imaging image encryption method based on chaotic mapping and DNA coding
Veličković et al. Resilience of MDCS Watermarking Algorithm in Wireless Network Environment

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant